The convergence of Quantum with AI
Compilers, circuit optimisation, error mitigation, and the algorithms themselves — including quantum machine learning and the hybrid classical-quantum stack.
227 stories
IBM and Qedma demonstrated quantum advantage in simulating the Floquet Ising model on a superconducting quantum processor, using Qedma's error mitigation to extract accurate dynamics beyond classical verification.
This could catalyze adoption of quantum simulation for short-time dynamics in materials science, as error mitigation proves sufficient to extract physically meaningful results on near-term devices.
Researchers at Nanyang Technological University demonstrated an AI method that significantly reduces the amount of measurement data required to characterize scalable quantum systems. The approach uses machine learning to infer system properties from fewer measurements, addressing a key bottleneck in quantum device calibration.
This could accelerate the calibration and tuning of quantum processors with increasing qubit counts, enabling faster development cycles for NISQ devices and early fault-tolerant systems.
IBM will install a Quantum System Two, its latest superconducting system with the Nighthawk processor, at the Swiss National Supercomputing Centre by the end of 2026. The installation is part of a defence-offset agreement with armasuisse and anchors a new innovation hub at ETH Zurich run with Lockheed Martin. The partners plan joint research in quantum sensing, additive manufacturing, and AI-adjacent algorithms.
The Zurich hub could become an operational testbed where Lockheed Martin and Swiss researchers run defence-relevant quantum sensing and additive-manufacturing algorithms on IBM's Nighthawk processor without moving data or hardware across US export boundaries.
A preprint on arXiv describes a framework that uses a graph neural network and proximal policy optimization reinforcement learning to automatically find compact parameterized quantum circuits. These circuits are then used as surrogate models for data from two different semiconductor device types: power GaN high-electron-mobility transistors and logic nanowire field-effect transistors. The framework aims to provide a unified, physics-aware approach to device modeling rather than relying on hand-designed quantum ansatze.
The RL-driven circuit discovery could be reapplied to other semiconductor devices, such as SiC MOSFETs or advanced FinFETs, by retraining on their I-V or C-V datasets.
QuEra researchers have demonstrated a machine learning technique that reduces quantum error rates by analyzing syndrome data. The method uses a neural network decoder to interpret error syndromes more accurately than traditional lookup-table approaches. This was tested on QuEra's neutral atom quantum platform.
Integration of ML-based syndrome decoders into QuEra's operational stack within two years could lower logical error rates enough to run deeper circuits on early fault-tolerant devices.
Researchers at Foshan University introduced a ranking loss function for quantum architecture search that improves the selection of quantum circuits by better aligning with noisy device performance.
The ranking loss could enable QAS to produce circuits with higher fidelity on current noisy quantum hardware, accelerating the deployment of quantum machine learning models.
Researchers have demonstrated that AI can learn to automatically construct quantum circuits for simulating drug molecules, drastically reducing the time required compared to manual design. The AI model was trained to generate circuits optimized for specific molecular properties, cutting design time by orders of magnitude.
AI-driven circuit generation could accelerate the practical use of quantum computers in drug discovery, making it feasible to screen molecular candidates on near-term devices within two years.